Sizing & selection
A tubular selection runs in a fixed order. The largest particle sets the bore, since a full-bore tube passes pieces up to about half its inner diameter. Bore and flow set the velocity. The duty comes from Q = mass flow × specific heat × temperature change, the temperature difference from the LMTD of a true counterflow layout, and the overall coefficient U mostly from the product side. Area divided by surface per foot gives tube length, and length divided into standard sections gives the unit. Pressure drop then checks the result, and a hold tube for pasteurizing duty is sized separately on time, not area.
The product film controls the answer. A 316L wall conducts about 9.4 Btu/h-ft-°F and is thin, so it adds little resistance; water in the annulus gives a high film coefficient. A viscous puree in laminar flow gives a low one, so U on thick product commonly runs well under half the water-to-water figure, and a unit for a sauce is several times longer than one for juice at the same duty. Viscosity also changes with temperature: a puree can be ten times thicker at the cold end than the hot end, which is why one viscosity figure is not enough.
The worked example below shows the arithmetic on a simple heater.
Sizing assumes a 316L product tube with a 304L or 316L annulus; the grade barely moves the arithmetic, but salt, chloride and wall temperature decide which goes on the order. Grade limits are set out on Material & Construction.
Pieces up to about half the inner diameter pass a full-bore tube; Q = mass flow × specific heat × temperature change sets the duty, and the product film, not the 316L wall, sets most of U.
Take 20 gpm of a thin sauce near water density, about 10,000 lb/h, with a specific heat of 0.9 Btu/lb-°F, heated from 40 to 180°F. Duty: 10,000 × 0.9 × 140 = 1,260,000 Btu/h. Hot water enters the annulus at 200°F and 70 gpm; by the 500 rule it gives up 1,260,000 / (70 × 500) = 36°F and leaves at 164°F. In counterflow the end differences are 20°F (200 against 180) and 124°F (164 against 40), so LMTD = (124 - 20) / ln(124/20) = 104 / 1.82, about 57°F.
Assume a clean U of 120 Btu/h-ft2-°F for this product. A fouling factor of 0.001 h-ft2-°F/Btu gives a design U of 1 / (1/120 + 0.001), about 107. Area = 1,260,000 / (107 × 57), about 207 ft2. A 2 in. OD product tube carries about 0.52 ft2 per foot, so the heater needs about 395 ft of tube: twenty 20 ft sections give 400 ft and about 209 ft2. At 20 gpm the velocity in a 2 in. tube with 0.065 in. wall is about 2.3 ft/s, and the bore passes pieces near 0.9 in. The example is illustrative; the real selection is run on rating software against the measured rheology, and the quotation, not this arithmetic, states the performance.
The same 207 ft2 can be built several ways. A single 2 in. tube suits a sauce with pieces; a multi-tube layout with 3–7 smaller inner tubes gives more surface per section for a thin liquid with no particles; a corrugated tube raises turbulence in viscous product, lifts U and slows fouling, at the cost of pressure drop and a profile that smaller pieces can catch on. Regeneration, where hot product preheats cold product, cuts the utility load and adds sections of its own.
Pressure drop is where viscous selections go wrong. In laminar flow it rises in direct proportion to viscosity and length, so the 400 ft that suits the duty can push the inlet pressure past what the pump delivers or the 150–300 psig product-side rating allows. The fixes are a larger bore, a corrugated tube to shorten the unit, or splitting the duty between sections that run at different velocities. Steam in place of hot water shortens the unit too, but lifts the wall temperature.
Rheology and particle data decide a tubular unit more than the duty figure does. A product sample settles most questions faster than estimates.
Call and talk it through with an engineer: 1-805-484-2992
Because it carries the product a plate unit cannot. One large bore with no contact points gives a low product film coefficient, especially in laminar flow, and the surface per foot of tube is modest. A plate pack reaches higher U through narrow, turbulent channels, which is exactly what pieces and pulp would plug. The length buys passage of solids and a crevice-free path.
It raises the LMTD, so the unit is shorter, but 15 psig steam condenses at 250°F and puts the tube wall close to that. Sugar, starch and protein products burn on at those wall temperatures, and on salted product a hot 316L wall above 140°F carries a chloride stress cracking risk. Pressurized hot water lets us cap the wall temperature, which is why most product heaters use it.
Send a sample or the recipe. Without a measured figure we assume conservatively, and the unit grows to match the assumption. Guessing low is the expensive mistake: the heater comes out short, misses its outlet temperature on the thick batches, and the pressure drop runs higher than the pump was chosen for. Two viscosity readings, cold and hot, remove most of the uncertainty.
A fouling factor of 0.001–0.002 h-ft2-°F/Btu is common on the product side, as in the example. Products that foul hard at heating temperatures, such as egg, high-protein dairy and starch sauces, are handled by run length and the cleaning interval more than by added surface, because extra length also adds pressure drop and residence time.
Brine, salted sauce at cooking heat and chloride effluent push past 316L. How 2205 duplex and 6-moly product tubes are chosen, welded and rated.
+ Learn More
Ra on the bore of a long product tube is set at the mill, at the weld and at the bend. What to specify for sticky, viscous and aseptic products.
+ Learn More
Why 316L is the standard product tube in a stainless double pipe unit, when the annulus can drop to 304L, and the salt and heat that move it to duplex.
+ Learn More